Hi , I'm Waris Radji
PhD Student in Deep/Theoretical Reinforcement Learning

profile picture

Hello , I'm Waris Radji
PhD Student in Deep/Theoretical Reinforcement Learning

Age
24 years
Location
France
Phone
+33 6 52 81 11 65
E-mail
Website
Interests
Street Lifting, Competitive Programming, Evolutionary Genetics
Software Engineer & Researcher with 3+ years of industry experience in AI system design. Combines practical software engineering expertise with theoretical foundations in deep reinforcement learning, bringing a unique interdisciplinary perspective that bridges RL research with scalable implementation.

WORK EXPERIENCE

Inria - Team Scool logo.
Jan 2023 - Now

Inria - Team Scool

PHD Student • supd. by Odalric-Ambrym Maillard • Lille

Bringing theoretical rigor to deep Reinforcement Learning exploration in high-dimensional stochastic environments, leveraging multi-armed bandits theory for principled yet scalable solutions.

Pytorch Pytorch logo
Python Python logo
Lunatech | Scala Center logo.
Jun 2022 - Aug 2022

Lunatech | Scala Center

OPEN-SOURCE DEVELOPER • Internship supd. by Chris Kipp • Rotterdam

Working on Scala 3 support and adding the call hierarchy feature into the open-source tool Metals, the Scala language server. I have been involved in the development of other minor features and in fixing many issues.

Scalameta Scalameta logo
Scala Scala logo
CNRS - Laboratoire d'Océanographie et du Climat logo.
Apr 2020 - Aug 2020

CNRS - Laboratoire d'Océanographie et du Climat

SOFTWARE ENGINEER • Internship supd. by Dimitry Khvorostyanov • Paris

Development of new functionalities for L-SAT, a web platform for collaborative analysis and exploration of satellite data. Minor changes to existing data analysis codes.

RabbitMQ RabbitMQ logo
ChartJS ChartJS logo
Javascript Javascript logo
Postgresql Postgresql logo
Matlab Matlab logo
Django Django logo
Python Python logo
CNRS - Northern Paris Computer Science Lab logo.
Jul 2019 - Aug 2019

CNRS - Northern Paris Computer Science Lab

MACHINE LEARNING ENGINEER • Associate • Paris

Implementation of a bio-inspired and clustering algorithm (Ant-Tree) in Scala to integrate it into the Clustering4Ever library. Use of data visualization libraries and Apache Spark to write ML notebooks.

Jupyter Jupyter logo
Maplotlib Maplotlib logo
Spark Spark logo
Python Python logo
Scala Scala logo

SKILLS

Programming
Scala, Python, Java, Haskell, JS/TS, C, C++, Kotlin
Data Engineering
Spark, Kafka, Google Cloud Platform, (No)SQL
Machine Learning
JAX, Pytorch, Scikit-Learn, TensorFlow, Deeplearning4j
Data Visualization
D3.js, Plotly, ChartJS, Processing, Matplotlib, ggplot
Web Framework
React, Play!, Flask, Django, Angular
Soft Skills
Adaptability, Creativity, Problem-Solving, Attention to Detail

EDUCATION

 logo
2022 -2023

Numerics Graduate Program - Optimisation and Decision

University of Bordeaux Graduate Research School

Combinatorial Optimization
Mathematical Modelling and Programming
Constraint Programming
Energy considerations in deep learning
 logo
2020 - 2023

Engineer's degree in Computer Science - Artificial Intelligence Major - PhD track

ENSEIRB-MATMECA - Bordeaux Graduate School in Electronics, Computer Science, Telecommunications, Mathematics and Mechanics

Probability
Formal Methods
Automata Theory
Compiler Theory
Distributed Computing
Numerical Methods
System Programming
Computability Theory
Machine Learning
Reinforcement Learning
Deep Learning
Video Analysis
Computer Vision
Search Algorithms
Knowledge Representation
 logo
2018 - 2020

Technical college degree in Computer Science

Université Sorbonne Paris Nord (DUT Informatique)

OOP
Database
HCI
Web Programming
Agile Methodology
Linear Algebra
Graph theory
Algorithms
Analysis
Differential Equation
Operational Research
Computer Network

ACHIEVEMENTS & PROJECTS

2022-2023

Evolutionary DIAYN

APPRENTICE RESEARCHER

Explored the benefits and limitations of applying DIAYN, a unsupervised reinforcement learning algorithm for skills discovery, on top of Evolutionary Strategies.

2022-2023

Reinforcement learning on symbolic data

APPRENTICE RESEARCHER

In collaboration with the INRIA Mnemosyne team, development of a variant of Q-Learning for use with symbolic data, allowing the incorporation of structured knowledge in the form of qualitative or quantitative data.

2022

Universal adversarial perturbations

APPRENTICE RESEARCHER

Reproduction of an experiment from the paper "Universal adversarial perturbations" (accepted at CVPR 2017) consisting in proving the efficiency of universal perturbations on the VGG11 architecture.

2021-2022

Eirb'IA

CO-FOUNDER (with Hector Piteau)

Creation and management of the Artificial Intelligence Club of ENSEIRB-MATMECA through workshops, conferences, debates, competitions and paper popularization.

2020-2022

Pokémon Studio

LEAD BACK-END DEVELOPER

Creation of Pokemon Studio, a tool developed in React to create its Pokemon game project more simply. The goal is to replace Ruby Host, the Database editor, and to allow more relevance by leaving RPG Maker XP.

2019-2021

Clustering4Ever

CONTRIBUTOR

Implement various ML algorithms into Clustering4Ever, a Big Data Clustering Library gathering clustering, unsupervised algorithms, and quality indices, written in Scala and Spark.